|
Abstract
|
Inthispaper,afullNesterov–Toddstepinfeasibleinterior-pointmethodfor
solvingsemidefiniteoptimizationproblemsbasedonanewkernelfunctionisanalyzed.
In each iteration, the algorithm involves a feasibility step and several centrality steps.
The centrality step is focused on Nesterov–Todd search directions, while we used
a kernel function with trigonometric barrier term to induce the feasibility step. The
complexityresultcoincideswiththebest-knowniterationboundforinfeasibleinterior-
point methods.
|